Biochimica et Biophysica Acta 1773 (2007) 254 – 263
www.elsevier.com/locate/bbamcr
Neuroprotective effect of TNFα against the β-amyloid neurotoxicity
mediated by CDK5 kinase
Daniel I. Orellana a , Rodrigo A. Quintanilla a , Ricardo B. Maccioni a,b,⁎
a
Laboratory of Cellular, Molecular Biology and Neurosciences, Faculty of Sciences, Universidad de Chile; Las Encinas 3370, Ñuñoa, Santiago, Chile
b
Department of Neurological Sciences, Faculty of Medicine, Universidad de Chile, Av. Salvador 486, Providencia, Santiago, Chile
Received 15 June 2006; received in revised form 2 October 2006; accepted 16 October 2006
Available online 21 October 2006
Abstract
The tumor necrosis factor alpha (TNFα) plays a dual role in producing either neurodegeneration or neuroprotection in the central nervous
system. Despite that TNFα was initially described as a cell death inductor, neuroprotective effects against cell death induced by several neurotoxic
insults have been reported. Tau hyperphosphorylation and neuronal death found in Alzheimer disease is mediated by deregulation of the cdk5/p35
complex induced by Aβ treatments. Since TNFα affects cdk5 activity, we investigated its possible protective role against the Aβ-induced
neurodegeneration, as mediated by cdk5. TNFα pretreatments significantly reduced the hippocampal neuronal cell death induced by the effects of
Aβ42 peptide. In addition, this pretreatment reduced the increase in the activity of cdk5 induced by Aβ42 in primary neurons. Next, we
investigated the Alzheimer type phosphorylation of tau protein induced by Aβ42. We observed that the pretreatment of neurons with TNFα
reduces tau hyperphosphorylation. Taken together, these results define a novel neuroprotective effect of TNFα in preventing neuronal cell death
and cdk5-dependent tau hyperphosphorylation. This phenomenon, taken together with other previous findings, suggests that the inflammatory
response due to Aβ peptide plays a key role in the development of Alzheimer etiopathogenesis.
© 2006 Elsevier B.V. All rights reserved.
Keywords: TNFα; Cdk5; Aβ; Tau; Neurodegeneration; Alzheimer disease
1. Introduction
TNFα is a pro-inflammatory cytokine produced by many cell
types. At the level of the central nervous system (CNS), TNFα can
be synthesized and released by astrocytes, microglia, and some
neuronal cell types [1,2]. TNFα also plays a key role in modulating normal brain physiology [3]. In addition, it has been shown
that TNFα is necessary for the synaptic plasticity in cortical
neurons [4,5]. Considerable studies have been made in understanding the molecular mechanism that mediates TNFα duality as
mediator of neurodegeneration or neuroprotection. TNFα exerts
its biological functions through the action of two main receptors,
TNF-R1 and TNF-R2, both of them induces NF-κB pathway
activation [6]. The activation of NF-κB by TNFα is necessary for
⁎ Corresponding author. Laboratory of Cellular, Molecular Biology and
Neurosciences, Faculty of Sciences, Universidad de Chile; Las Encinas 3370,
Ñuñoa, Santiago, Chile.
E-mail address: [email protected] (R.B. Maccioni).
0167-4889/$ - see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.bbamcr.2006.10.010
neuronal survival [7,8]. Additionally, a model mouse that lack of
TNFα receptors is more susceptible than wild-type animals to
brain injury and neuronal damage [9,10,11]. Binding of TNFα to
its receptor and the activation of NF-κB pathway seems to be
critical in the dual role of this cytokine [12,6].
Currently, it is well accepted that inflammatory mechanism is
involved in the pathogenesis of Alzheimer disease (AD). In AD
cytokines such as IL-6, IL-1β and TNFα are secreted by
microglial cells, astrocytes and/or neuronal cells, and can
induce the synthesis of amyloid precursor protein (APP) [13].
On the other hand, APP or the amyloid beta peptide (Aβ) can
induce the expression of IL-6, IL-1β or TNFα in astrocytes and
microglial cells in culture [14,15].
The cdk5/p35 complex plays a pivotal role in the development of nervous system, as it has been shown through
engineered targeted mutations of cdk5, p35, and p39 genes
producing cortical migration defects [16,17]. The function of
this enzyme, a member of the proline-directed protein kinases
(PDPKs) family that also includes the glycogen synthase kinase
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
3β (gsk3β), is regulated by the neurospecific activators p35
[18,19] and p39 [20]. Overactivation of PDPKs, such as cdk5
and gsk3β, has been implicated in the aberrant phosphorylation
of tau at proline-directed sites implicated in the ethiopathogenesis of Alzheimer disease [21,22].
It is known that an increase in the cdk5 activity is an important factor in neurodegeneration [23,24]. We have previously
demonstrated that both Aβ and IL-6 treatments of hippocampal
neurons induce an increase in the cdk5 activity [25,26]. The
cdk5-mediated neurotoxicity is, at least in part, due to an
increase in tau hyperphosphorylation [25]. Additional studies
have proposed that cdk5 inhibits the neuroprotective effect
mediated by the transcription factor MEF2 [27,28]. Moreover,
evidence suggests that TNFα modulates the expression of MEF2
in mouse soleus muscle [29]. These observations indicate a
possible common activation pathway.
Moreover, pretreatment of primary neurons with TNFα
protected neurons against Aβ toxicity by suppressing accumulation of ROS and cytosolic Ca2+ increases [12]. Until now, the
neuroprotective effect of TNFα and the possible role of cdk5 in
this event are unknown. In the present study, we show that
pretreatment of hippocampal neurons with TNFα prevents the
neurodegeneration induced by Aβ42 by inhibiting the cdk5
activity and decreasing their protein levels in hippocampal
neurons. Additionally, the pretreatment of TNFα prevented tau
hyperphosphorylation induced by Aβ42, and decreased the
cortical neurons aggregation, both processes governed by the
cdk5. Even though p35, the cdk5 activator, is involved in cdk5
function, no significant changes in protein expression were
evidenced. Based on these findings, we suggest that in hippocampal neurons, TNFα protects cells against Aβ neurotoxicity, by modulating cdk5 activity. These events may contribute
to the whole neuroprotective effect exerted by TNFα in AD.
2. Materials and methods
2.1. Materials
Human recombinant TNFα was obtained from US Biologicals. Aβ42 was
purchased from Global Peptide Services. Cdk5 monoclonal antibody (mAb) J3
and polyclonal antibody C8; p35 mAb N20 were obtained from Santa Cruz
Biotechnology; AT8, mAb was from Innogenetics (Belgium); Tau-1 mAb and
Tau5 mAb were donated by Dr. Lester Binder. Dr. Peter Davis generously
donated PHF1 mAb. MitotoCapture™ apoptosis detection kit and roscovitine
were purchased from Calbiochem (San Diego, La Joya). Tissue culture reagents
were obtained from Gibco BRL.
255
aliquots of Aβ peptide in dimethylsulfoxide (DMSO) at 15 mg/ml (3.5 mM).
Aliquots of peptide stock solution (70 nmol in 20 μl DMSO) were added to
sterilize MilliQ H2O. The solution was incubated for 3 days, after which the Aβ
fibrils were concentrated by centrifugation (14 000 rpm for 30 min) to separate
the soluble supernatant containing oligomers of the peptide from the insoluble
fibrillary (pellet) form of the peptide. Soluble oligomeric forms of the peptide
were resuspended in Neurobasal/B27 media and stored at − 20 °C. This soluble
Aβ42 was added to the culture medium at a final concentration of 10 μM by
24 h, a concentration that has been shown to cause early events of neurotoxicity
in cultured hippocampal neurons [21]. These conditions of a non-detergentcontaining extraction buffer allow the formation of a soluble Aβ peptide
fraction containing many oligomeric species, which is more toxic than fibrillary
Aβ [30].
2.4. Neuronal aggregation assay
Dissociated rat embryonic cortical neurons (E18) were washed with Hank
1× and resuspended in plating media. The neurons were plating in media with
1 mM EGTA, 1 mM CaCl2, and TNFα (5 ng/ml). Then, the neuronal
aggregation assay was carried out as described in [31] with minor modifications.
Briefly, the cells were incubated at 80 rpm in a 37 °C floor shaker. After 15 min
at 37 °C, aggregates were fixed with 4% paraformaldehyde (10 min) and
permeabilized with PBS containing 0.2% Triton X-100 (10 min). Coverslips
were rinsed with PBS followed by H2O. Cresyl violet stain was applied for
5 min and the coverslips mounted with PBS containing 50% glycerol. The
surface of 10 viable aggregates (5 or more cells) for each condition was
measured using the software Axiovision 3.5 (Carl Zeiss, Germany).
2.5. Cell viability
Cell viability was evaluated using the MTT assay previously described at
[26]. Briefly, neurons were seeded in polylysine-coated 96-well plates at
2.0 × 104 cells/100 μl per well in Neurobasal/B27 medium without phenol red.
Then, neurons were treated with TNFα 5 ng/ml, Aβ42 10 μM, or both. The
results were represented in the graphs as a percent of control.
2.6. Mitochondrial viability
MitoCapture™ Apoptosis Detection Kit was used in order to distinguish
between healthy and apoptotic cells by detecting the changes in the mitochondrial
membrane potential. In healthy cells, MitoCapture™ reagent accumulates and
aggregates in the mitochondria, emitting a bright red fluorescence. In apoptotic
cells, this reagent cannot aggregate in the mitochondria due to the altered
mitochondrial membrane potential, and thus it remains in the cytoplasm in its
monomeric form and emits a green fluorescence. Fluorescence changes and
images were acquired and analyzed with a Zeiss LSM 510 META confocal
microscope (Carl Zeiss, Germany). The filters used were BP 505–530 and HFT
488 for green, and LP 560 and HFT488/543 for red. Laser excitations were
488 nm 5% and 543 nm 80%. Quantification of 20 cells per condition was
measured using the software LSM 510 META. The results were represented in
the graphs as a fluorescence ratio between red/green channels, respectively.
2.7. Immunodetection assays
2.2. Primary rat cultures of hippocampal neurons
Primary rat hippocampal neuronal cultures were established from the
hippocampus of 18-day-old Sprague–Dawley rat fetuses according with methods
described previously [26]. Neurons were seeded onto 0.1% polylysine-coated 60mm plastic Petri dishes (1.0 × 106cell/60-mm culture dish) for biochemical
experiments. TNFα at a concentration of 5 ng/ml was added to hippocampal
neurons culture at 4 days in vitro (4-DIV), which continued in culture for additional
48 h. Sister cultures that did not receive TNFα treatment were used as controls.
The immunological reactivity of hippocampal cell extracts treated with
cytokines and the untreated controls was assayed by Western blots techniques,
as previously described by [26] using antibodies that recognize Alzheimer
phosphoepitopes on tau (AT8, PHF-1), cdk5 protein, p35 protein, and with the
antibodies Tau1 and Tau5. Quantification of blots was carried out by scanning
the photographic films of nitrocellulose membranes, by using the Kodak digital
Science densitometry program from Kodak. Molecular size markers were from
Bio-Rad Laboratories.
2.3. Amyloid-β preparation
2.8. Immunoprecipitation and cdk5 activity assays
Synthetic Aβ42 peptide solutions (>70% HPLC) were prepared as described
previously [21]. Briefly, stock solutions were prepared by dissolving lyophilized
Primary hippocampal cultured neurons were plated at 1 × 106 cells/cm2 on
polylysine-coated 60-mm dishes. After treatments, cells were lysed in RIPA
256
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
buffer plus protease inhibitors. 200 μg of total protein was used for
immunoprecipitation with an anti-cdk5 J3 monoclonal antibody used at a final
dilution of 1:50. Then, the antigen–antibody complex was captured with either
agarose–proteins A or G. Then, in vitro kinase assay was carried out as
described by [26]. Quantification was autoradiography with a Molecular Imager
FX (Bio-Rad).
2.9. RT-PCR assays
Total RNA was extracted from the control, as well as from either, TNFα,
Aβ42 or Aβ + TNFα-treated hippocampal cell cultures using the TRIZOL
(Gibco BRL) reagent. Then RT-PCR was carried out as described [26]. PCR
amplification of samples was performed with specific primers against cdk5
and actin. Cdk5 primers were forward: 5′ GCATTGAGTTTGGGCACGACA
3′, reverse: 5′AAAACCGGGAAACCCATGAGA 3′; β-actin, forward: 5′
TCTACAATGAGCTGCGTGTG 3′, reverse:5′TACATGGCTGGGGTGTTGAA 3′.
2.10. Statistical analyses
One-tailed Student's t test for regression analyses was performed with the
SigmaPlot software (Jandel Scientific, Corte Madera, CA). Data are presented as
means ± SEM (n). n = number of experiments. Statistical significance was
considered with p values equal or lower than 0.05 (p ≤ 0.05).
3. Results
3.1. TNFα does not affect the cell viability of rat hippocampal
neurons and protects them from the neuronal cell death
induced by Aβ42
Our first goal was to define the effect of relatively low
concentrations of TNFα on the viability of cell cultures. Thus,
Fig. 1. TNFα protects from the mitochondrial viability loss induced by Aβ peptide in hippocampal neurons. (A) Cultures of hippocampal neurons were treated with
TNFα (5 ng/ml), Aβ42 (10 μM) or Aβ42 plus TNFα pretreatment. Fluorescence changes were measured by using MitoCapture kit (see Materials and methods). Neither
fluorescence variations nor morphological change was observed between control cells and cells treated with TNFα. (B) Quantification of the fluorescence changes
measured with MitoCapture™ kit (20 neuronal cells, n = 4). No significant differences in the mitochondrial viability after treatment with TNFα were evident, between
control neurons and treated with TNFα. Nevertheless, a significant reduction in the mitochondrial viability was observed in the cultures treated with Aβ42 which was
reverted by pretreatment with TNF. The asterisk denotes statistical differences (p < 0.05).
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
hippocampal neurons were treated with TNFα (5 ng/ml by 48 h).
The direct administration of TNFα did not affect cell viability as
determined by MitoCapture apoptosis kit and MTT reduction
assays (see Figs. 1 and 2). No fluorescence variations neither
morphological change were observed, when control cells were
compared with hippocampal neurons treated with TNFα (Fig.
257
1A). Quantification of fluorescence, expressed as a ratio
between red and green fluorescence, demonstrated that TNFα
did not induce mitochondrial potential loss [control: 1.633
± 0.044; TNFα: 1.719 ± 0.16; n = 3] (Fig. 1B). Consistently,
MTT reduction assays also indicated that TNFα treatments had
no effect on neuronal cell death (Fig. 2A) [control: 99.75
± 0.88%; TNFα: 88.7 ± 1.4%; n = 5]. Although TNFα alone is
apparently toxic as observed in the MTT data in Fig. 2A, these
results are agreement with previous observations [12], which
showed that a bigger TNFα concentration (100 ng/ml) induced
a similar neuronal cell death.
On the basis of these results, it was important to determine
the protective role of TNFα in hippocampal neurons that were
exposed to Aβ42 peptide. Thus, we carried out similar MTT and
mitochondrial viability studies. In this study, hippocampal
neurons were pretreated with TNFα (5 ng/ml by 48 h) and then
treated with Aβ42 (10 μM) for additional 24 h. The effect on cell
viability by Aβ42 alone was also measured. A severe decrease in
the red fluorescence and an increase in the green fluorescence
can be observed in the neurons treated with Aβ42, indicating
that Aβ42 is able to disrupt the mitochondrial potential in
hippocampal neurons (Fig. 1A). However, pretreatments with
TNFα were able to reverse the decrease in the red fluorescence
induced by Aβ42 and remains mitochondrial potential unaltered
(Fig. 1A and B). Quantification of fluorescence showed that
Aβ42 induced a decrease of 2.6 times in the fluorescence ratio
regarding control cells, and the TNFα treatment was able to
reverse cell death induced by Aβ42 treatment (Fig. 1B) [Aβ:
0.588 ± 0.03; Aβ + TNFα: 1.229 ± 0.06; n = 3, p < 0.001]. Consistently, MTT assays show that Aβ42 treatments reduce cell
viability (Fig. 2A) [control: 99.75 ± 0.88; Aβ: 62.7 ± 3.3%;
n = 5], while pretreatment with TNFα reverted the neurodegenerative effect of Aβ on hippocampal neurons (Fig. 2A) [Aβ:
62.7 ± 3.3%; Aβ + TNFα: 95.8 ± 1.4%; n = 5]. As a control,
hippocampal neurons treated with 10 μM roscovitine (a cdk5
inhibitor) did not show neuronal cell death. Then, the cdk5
inhibition by roscovitine, applied in the presence of Aβ,
showed to protect from cell death induced by the amyloid
peptide (Fig. 2D, Aβ + R). Interestingly, roscovitine protects
from Aβ-induced neuronal death, in the same fashion that
TNFα treatment does (Fig. 2A). Moreover, when hippocampal
neurons were preincubated under both conditions (TNFα and
roscovitine), the total protection reach out equivalent percentage (Fig. 2A). These observations indicated that TNFα and
roscovitine could induce protection against neuronal death by
using a common pathway.
Fig. 2. TNFα effects on cells viability of rat hippocampal neurons. Cell viability
was evaluated using the MTT reduction assay. (A) MTT assay (n = 5) shows that
TNFα does not affect the hippocampal neurons viability and the neuronal cell
death induced by Aβ was prevented in cells pretreated with this cytokine.
Roscovitine (10 μM), a cdk5 inhibitor, did not show neuronal cell death, and
protects from the cell death induced by Aβ42, as well as TNFα. The asterisk
denotes statistical differences (p < 0.05). (B) Constant concentration of TNFα
(5 ng/ml) still showed a protective effect against increasing Aβ42 concentrations.
(C) Increasing concentrations of TNFα did not induce an increase in the
hippocampal viability and still protects against Aβ42 neurotoxicity.
258
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
In order to determinate whether TNFα still has a protective
effect against increasing concentrations of Aβ42, we varied the
peptide concentration from 10 to 20 μM maintaining stable the
TNFα concentration (5 ng/ml). Under these conditions, TNFα
still showed a protective effect against the highest Aβ42
concentrations (Fig. 2B). In addition, we varied the concentration of TNFα from 5 to 100 ng/ml but maintaining stable the
Aβ42 concentration (10 μM). Increasing concentrations of
TNFα were not able to induce an increase in the hippocampal
viability, and these bigger concentrations still showed a
protective effect against Aβ42 (Fig. 2C). These results suggest
that TNFα prevents neuronal cell death as induced by Aβ in rat
hippocampal neurons. The following experiments were carried
out to explore the molecular mechanisms involved in these
cellular events.
3.2. TNFα induces a decrease in the activity of cdk5 in rat
hippocampal neurons
Although, it is known that pretreatment of hippocampal
neurons with TNFα prevents neuronal cell death in Aβ-treated
cells [12], the role of cdk5/p35 complex in the neuroprotection
induced by TNFα is a novel event. With the aim of exploring
this issue, we studied the levels of cdk5 protein in hippocampal
neurons that were treated with TNFα alone or pretreated with
TNFα and then exposed to Aβ42 peptide. A significant
reduction of the cdk5 protein levels was observed in hippocampal neurons after TNFα pretreatments (Fig. 3A) [control:
106.72 ± 6.63; TNFα: 69.13 ± 12.22; p < 0.05]. Those hippocampal neurons pretreated with TNFα, in the presence of Aβ42
also showed a significant reduction in the cdk5 proteins levels
(Fig. 3A) [Aβ: 115.48 ± 17.54; Aβ + TNFα: 73.77 ± 16.09;
p < 0.05]. Additionally, the Aβ42 treatment did not induce any
particular changes in the cdk5 protein levels regarding the
control (Fig. 3A) [control: 106.72 ± 6.63; Aβ: 115.48 ± 17.54;
p = 0.236]. Due to the decrease in the cdk5 protein levels, we
decide to study whether there are changes in the cdk5 mRNA
levels in response to TNFα treatment. We did not observe
significant changes in the cdk5 mRNA levels (Fig. 3B). This
result suggests that TNFα treatment decrease the cdk5 protein
levels, and that this deregulation could involve some posttranslational modifications.
Since, we found a reduction in the cdk5 protein levels, the
next step was to study the cdk5 activity. Interestingly, TNFα
pretreatment significantly reduced the pathological increase of
cdk5 activity by the effects of Aβ42 (Fig. 4A) [Aβ: 150.63 ±
10.18; Aβ+TNFα: 112.55 ± 9.73; p < 0.01]. Aβ42 induced an
increase in the cdk5 activity with respect to control cells [control:
121.07 ± 6.63; Aβ: 150.63 ± 10.18; p = 0.02], in agreement with
our previous observations [25]. Significant reduction of the cdk5
activity was also observed in TNFα-treated neurons without Aβ
[control: 106.72 ± 6.63; TNFα: 97.93 ± 2.88; p = 0.02]. Roscovitine was used as a control to inhibit the cdk5 activity increased
by Aβ42 treatment, and prevents Aβ-induced cdk5 activity, as
well as TNFα treatment (data not shown).
Consistent with the fact that the cdk5 activity depends of its
specific activators p35 and p39 [32,33], we have observed that
Fig. 3. TNFα induces a decrease in the cdk5 protein levels in rat hippocampal
neurons. (A) Western blot assay (n = 5) of hippocampal neurons untreated and
treated with TNFα (5 ng/ml by 48 h), showing that TNFα induces a decrease in
the cdk5 protein levels, analyzed with a mAb against cdk5 (J3) (n = 5). Protein
levels were normalized with the endogenous actin levels. (B) RT-PCR of
hippocampal neurons shows that TNFα did not induce significant changes in the
DNA fragment of 924 bp corresponding to cdk5 as compared with untreated
control. RT-PCR amplification was normalized with endogenous β-actin
transcript and not significant differences was observed for mRNA (n = 3).
IL-6 induces deregulation of the cdk5 activity increased the p35
levels [26]. Thus, we investigated the effect of TNFα in the p35
protein levels. In these studies, although p35 was present, TNFα
did not induce significant changes on the p35 levels between
control and TNFα-treated neurons (Fig. 4B) [control: 89.58 ±
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
259
affect this process, neuronal aggregation assays were carried
out. This assay showed that viable aggregates were formed in
the presence of TNFα. Besides, it was observed that 1 mM
CaCl2 used as positive control favored aggregates, as opposed
by negative effects in the presence of EGTA (Fig. 5A).
Quantification of two-dimensional surface area of individual
aggregates revealed that TNFα led to an increased aggregation
of rat cortical neurons (Fig. 5B) [control: 61.53 ± 9.16; TNFα:
544.57 ± 66.28; p < 0.01]. However, the effect in the aggregation
phenomenon did not observed in presence of TNFα plus EGTA.
This suggests that TNFα itself does not induce a calciumstimulated neuronal aggregation suggesting another mechanism
of action. These results are in agreement with Kwon et al. [31],
indicating an indirect role of TNFα facilitating the neuronal
aggregation, perhaps modifying the cdk5 activity.
3.3. TNFα decrease the Aβ42 induced Alzheimer type
phosphorylation of tau in rat hippocampal neurons
Fig. 4. TNFα induces a decrease in cdk5 activity in rat hippocampal neurons. (A)
In vitro kinase assay shows that TNFα induces a significant reduction in the cdk5
activity. Pretreatments with TNFα were able to reduce the increase in the cdk5
activity induced by Aβ42 peptide. In the lower panel we show a densitometric
analysis from 4 different experiments. (B) Western blot analysis of hippocampal
neurons shows no differences in the p35 levels, analyzed with a polyclonal
antibody against p35 (C20) (n = 4) between control and the different treatments.
The p35 protein levels were normalized with the endogenous actin levels.
10.68; TNFα: 73.25 ± 6.65; p = 0.09] in neurons exposed to Aβ,
and Aβ + TNFα (Fig. 4B) [Aβ: 78.69 ± 11.47; Aβ + TNFα:
54.93 ± 9.84; p = 0.06]. These observations discard a direct role
of p35 in the regulation of cdk5 activity by TNFα.
According to Kwon et al., one of the functions of cdk5 is to
regulate the adhesion process in cortical neurons [31]. In order
to evaluate whether TNFα modulating cdk5 activity was able to
In order to evaluate whether the reduction of cdk5 activity
induced by TNFα had and effect on the tau phosphorylation
patterns, we used antibodies that recognize hyperphosphorylated tau epitopes (Fig. 6). A significant reduction on the
hyperphosphorylated levels of tau recognized by AT8 antibody
was observed in the hippocampal neurons preincubated with
TNFα. Interestingly, pretreatments with TNFα also reduced the
AT8-tagged hyperphosphorylated tau that it was increased by
Aβ (Fig. 6) [control: 38.19 ± 4.4; TNFα: 29.32 ± 1.9; p ≤ 0.05;
and Aβ: 59.2 ± 9.8; Aβ + TNFα: 23.64 ± 10; p < 0.05]. Consistently, by using PHF1 antibody, TNFα also reduces the PHF1tagged hyperphosphorylated tau (Fig. 6) [control: 67.2 ± 10.95;
TNFα: 43.62 ± 7.3; p ≤ 0.05; and Aβ: 78.52 ± 8.82; Aβ+TNFα:
52.97 ± 5.94; p < 0.05]. On the other hand, the levels of
unphosphorylated tau, detected with the Tau1 antibody, were
increased in TNFα treatments (Fig. 6) [control: 44.73 ± 1.49;
TNFα: 52.49 ± 2.65; p < 0.05; and Aβ: 43.95 ± 1.43; Aβ+TNFα:
50.24 ± 2.68; p < 0.05]. As a control, no differences were
observed in the Tau5 levels (Fig. 6). These results indicate
that TNFα prevents Aβ-induced tau hyperphosphorylation, an
event that could be explained for an apparent decrease in cdk5
activity induced by TNFα. Additionally, TNFα regulating cdk5
activity can prevent Aβ-induced neurotoxicity in primary
hippocampal neurons, showing a novel neuroprotective role
of this cytokine in Alzheimer disease.
4. Discussion
In this study, we demonstrate that TNFα regulates the cdk5
activity, modulating the cdk5 protein levels. In addition, TNFα
pretreatment induced a significant reduction in the cdk5 activity
increased by Aβ42. Survival of hippocampal neurons treated
with TNFα was not compromised, thus allowing us to carry out
most of our studies under conditions that assure high cell
viability. Despite that MTT assay has shown a decrease in the
cell viability of 10% upon TNFα treatments, the percentage of
cell death is acceptable and allowed us to carry out all our
experiments under low conditions of cell damage. This result is
260
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
Fig. 5. TNF induces aggregation of rat embryonic cortical neurons. (A) Dissociated embryonic cortical neurons were aggregated in the presence of 1 mM EGTA, 1 mM
CaCl2, TNFα (5 ng/ml) or 1 mM CaCl2 with TNFα. Aggregates formed in the presence of 1 mM CaCl2 but not in 1 mM EGTA. In the presence of 1 mM CaCl2 and
TNFα, viable aggregates formed were larger. (B) Quantification of two-dimensional surface area of 10 aggregates (n = 3) is shown. The asterisk denotes statistical
differences (p < 0.05).
in agreement with previous data where TNFα protected hippocampal neurons from Aβ40 toxicity [12]. Thus, TNFα showed
a potent neuronal death protection in Aβ-treated neurons.
Nevertheless, in our studies we used soluble Aβ42 which is much
more susceptible to aggregation than Aβ40, and is much more
toxic than the other forms of Aβ peptide [30,34].
It has been shown that Aβ induces an increase in the cdk5
activity, and it induces phosphorylations of tau protein at
Alzheimer epitopes [21,35]. The use of inhibitors against cdk5
reduces the increase in the cdk5 activity and the tau hyperphosphorylation and apoptosis in neurons exposed to Aβ peptide
[21,36]. In regard with possible mechanisms, our results suggest
that TNFα could be acting by inhibiting cdk5, because it
significantly reduces the rise in cdk5 activity promoted by Aβ.
In addition, TNFα decreases tau hyperphosphorylation and
neuronal death.
It is known that inhibition of cdk5 kinase activity by using
the specific inhibitor roscovitine leads to the formation of larger
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
261
Fig. 6. TNFα prevents tau hyperphosphorylation induced by Aβ in hippocampal neurons. Western blot assay (n = 3) of hippocampal neurons untreated and treated with
TNFα (5 ng/ml by 48 h) or Aβ42 (10 μM by 24 h) with TNFα, showing that TNFα induces a decrease in the tau hyperphosphorylated levels, analyzed with a mAb
against AT8, PHF1. On the other hand, TNFα induces an increase in unphosphorylated Tau1 levels. As control the levels of tau were measured with Tau5. AT8
recognizes hyperphosphorylated tau epitopes in Ser202 and Thr205. PHF1 recognizes hyperphosphorylated tau epitopes in Ser396 and Thr404. Tau1 recognizes
unphosphorylated tau epitopes in Ser202 and Thr205. Tau5 recognizes a conformational epitope on tau in a phosphorylation-independent manner. Protein levels were
normalized with the Tau5 levels. The asterisk denotes statistical differences (p < 0.05).
aggregates of embryonic cortical neurons [31]. Here, we
reported that TNFα induces an equivalent effect on cortical
neurons aggregation. This result reinforces the inhibitory effect
of TNFα on the reduction of the cdk5 kinase activity as measured by in vitro kinase assay.
Even though these studies confirm previous results supporting TNFα protective role [12], our findings support a novel
effect of TNFα in promoting a neuroprotective action against
Aβ peptide, and particularly on the soluble set of oligomeric
forms of the amyloid peptide. In this paper, we describe that
TNFα treatment reduces cdk5 activity affecting protein levels
and activity. However, TNFα can also activate other signaling
pathways involved in neuroprotection and tau hyperphosphorylation, such as: MAPK, p38, and gsk3β [28,36]. This last
enzyme does not show any particular changes in their protein
levels and activation in hippocampal neurons exposed to TNFα
(data not shown). For this reason, and gathering previous
evidence [21,25,28,36], we postulate that TNFα could protect
hippocampal neurons by reducing the cdk5 activity and
preventing tau hyperphosphorylation and neuronal death.
Even though the activator p35 is present, no significant
changes in its protein expression were evidenced. Despite the
fact the cdk5 activity is also regulated by its activator p35 [32],
and that we were able to detect p35 levels present in our system,
TNFα pretreatment did not produce any significant changes in
the p35 protein levels, thus discarding its role in the TNFαmediated regulation of cdk5 activity. Previous data have
reported changes in the cdk5 activity in colon cancer cells
with a decrease in cdk5 protein expression and kinase activity
without significant changes in the p35 protein levels [37].
Additionally, there is evidence that shows that the promoter
region of cdk5 displays specific sites for the interaction with the
transcriptional factor AP-1 [38]. Moreover, TNFα can induce
their effects through AP-1 activation [6], so it is possible that
TNFα regulates the cdk5 protein levels and its activity by acting
on cdk5 post-translational modifications. In addition, some
studies have documented that cdk5 inhibits the neuroprotective
effect mediated by MEF2 [27, 28] and phosphorylates STAT3
regulating its transcriptional activity [39]. Consistently, it has
been shown that TNFα modulates the expression of MEF2 in the
mouse soleus muscle [29], and STAT3 in hepatic cells [40].
Therefore, it is also possible that TNFα could also modify MEF2
and STAT3 through a cdk5-dependent mechanism. However,
further studies are required to inquire on this hypothesis.
On the other hand, the use of specific antibodies against
hyperphosphorylated tau epitopes such as AT8 and PHF-1, and
262
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
antibodies that recognized the unphosphorylated tau isoforms
such as Tau1 allows us to visualize tau phosphorylating patterns
by proline-directed protein kinases (PDPKs), making cdk5 a
reasonable candidate for such an effect. Nevertheless, it is
known the role of phosphatases such as PP1, PP2A and PP2B in
the regulation of tau phosphorylation patterns [41]. It has been
demonstrated that TNFα induces the activation of protein
phosphatases [42]. Therefore, in our system, TNFα might also
be activating phosphatases and inducing tau dephosphorylations, an aspect that remains to be investigated in depth.
There are some studies which have shown TNFα neurotoxicity [43,44]. Apparently, in these cases the toxicity was mediated
by activated microglia cells from cortical cultures, which tend to
have a larger number of glial cells. The hippocampal cultures
used in our study are completely depleted of glia, since these
cultures were treated in the presence of Ara-C (10 mM). These
suggest that the neuroprotective effect of TNFα is not mediated
by glia.
In summary, our data suggest that TNFα induces a
neuroprotective effect in hippocampal neurons against Aβ42
neurotoxicity. These phenomena suggest that the inflammatory
response due to Aβ effects could play a key role in Alzheimer
disease etiopathogenesis. Taking together, the present data
provide a novel molecular mechanism involved in these
neuroprotective events of TNFα in preventing neuronal cell
death and tau hyperphosphorylation dependent of cdk5.
Acknowledgement
This research was supported by grants from FONDECYT
project 1050198 (to R.B.M.), and by support from the
International Center for Biomedicine (ICC). D.O. was a
predoctoral fellow of the biotechnology program, from the
Millennium Institute CBB.
References
[1] B. Cheng, S. Christakos, M.P. Mattson, Tumor necrosis factors protect
neurons against metabolic–excitotoxic insults and promote maintenance of
calcium homeostasis, Neuron 12 (1994) 139–153.
[2] R.T. Perry, J.S. Collins, H. Wiener, R. Acton, R.C. Go, The role of TNF
and its receptors in Alzheimer's disease, Neurobiol. Aging 22 (2001)
873–883.
[3] L. Vitkovic, J. Bockaert, C. Jacque, “Inflammatory” cytokines: neuromodulators in normal brain? J. Neurochem. 74 (2000) 457–471.
[4] E.C. Beattie, D. Stellwagen, W. Morishita, J.C. Bresnahan, B.K. Ha,
M. Von Zastrow, M.S. Beattie, R.C. Malenka, Control of synaptic strength
by glial TNFalpha, Science 295 (2002) 2282–2285.
[5] D. Stellwagen, R.C. Malenka, Synaptic scaling mediated by glial TNFalpha, Nature 440 (2006) 1054–1059.
[6] V. Baud, M. Karin, Signal transduction by tumor necrosis factor and its
relatives, Trends Cell Biol. 11 (2001) 372–377.
[7] B. Kaltschmidt, M. Uherek, H. Wellmann, B. Volk, C. Kaltschmidt,
Inhibition of NF-kappaB potentiates amyloid beta-mediated neuronal
apoptosis, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 9409–9414.
[8] P. Fernyhough, D.R. Smith, J. Schapansky, R. Van Der Ploeg, N.J.
Gardiner, C.W. Tweed, A. Kontos, L. Freeman, T.D. Purves-Tyson, GW.
Glazner, Activation of nuclear factor-kappaB via endogenous tumor
necrosis factor alpha regulates survival of axotomized adult sensory
neurons, J. Neurosci. 25 (2005) 1682–1690.
[9] A.J. Bruce, W. Boling, M.S. Kindy, J. Peschon, P.J. Kraemer, M.K.
Carpenter, F.W. Holtsberg, M.P. Mattson, Altered neuronal and microglial
responses to excitotoxic and ischemic brain injury in mice lacking TNF
receptors, Nat. Med. 2 (1996) 788–794.
[10] P.G. Sullivan, A.J. Bruce-Keller, A.G Rabchevsky, S. Christakos, D.K.
Clair, M.P. Mattson, S.W. Scheff, Exacerbation of damage and altered NFkappaB activation in mice lacking tumor necrosis factor receptors after
traumatic brain injury, J. Neurosci. 19 (1999) 6248–6256.
[11] C. Thompson, D. Gary, M. Mattson, A. Mackenzie, G.S. Robertson,
Kainic acid-induced naip expression in the hippocampus is blocked in
mice lacking TNF receptors, Brain Res. Mol. Brain Res. 123 (2004)
126–131.
[12] S.W. Barger, D. Horster, K. Furukawa, Y. Goodman, J. Krieglstein, M.P.
Mattson, Tumor necrosis factors alpha and beta protect neurons against
amyloid beta-peptide toxicity: evidence for involvement of a kappa Bbinding factor and attenuation of peroxide and Ca2+ accumulation, Proc.
Natl. Acad. Sci. U. S. A. 92 (1995) 9328–9332.
[13] T. Wyss-Coray, L. Mucke, Inflammation in neurodegenerative disease—A
double-edged sword, Neuron 35 (2002) 419–432.
[14] B.D. Gitter, L.M. Cox, R.E. Rydel, P.C. May, Amyloid beta peptide
potentiates cytokine secretion by interleukin-1 beta-activated human
astrocytoma cells, Proc. Natl. Acad. Sci. U. S. A. 92 (1995) 10738–10741.
[15] I. Blasko, M. Stampfer-Kountchev, P. Robatscher, R. Veerhuis, P.
Eikelenboom, B. Grubeck-Loebenstein, How chronic inflammation can
affect the brain and support the development of Alzheimer's disease in old
age: the role of microglia and astrocytes, Aging Cell 3 (2004) 169–176.
[16] T. Ohshima, J.M. Ward, C.G. Huh, G. Longenecker, H.C. Pant, R.O.
Brady, L.J. Martin, Targeted disruption of the cyclin-dependent kinase 5
gene results in abnormal corticogenesis, neuronal pathology and perinatal
death, Proc. Natl. Acad. Sci. U. S. A. 93 (1996) 11173–11178.
[17] T. Chae, Y.T. Kwon, R. Bronson, P. Dikkes, E. Li, Mice lacking p35, a
neuronal specific activator of cdk5, display cortical lamination defects,
seizures, and adult lethality, Neuron 18 (1997) 29–42.
[18] L.H. Tsai, I. Delalle, V.S. Caviness, T. Chae, p35 is a neural-specific
regulatory subunit of cyclin-dependent kinase 5, Nature 371 (1994)
419–423.
[19] G. Alvarez, J.R. Munoz-Montano, J. Satrustegui, J. Avila, E. Bogonez, J.
Diaz-Nido, Lithium protects cultured neurons against beta-amyloidinduced neurodegeneration, FEBS Lett. 453 (1999) 260–264.
[20] S. Humbert, R. Dhavan, L.H. Tsai, p39 activates cdk5 in neurons, and is
associated with the actin cytoskeleton, J. Cell Sci. 113 (Pt 6) (2000)
975–983.
[21] A. Alvarez, R. Toro, A. Caceres, R.B. Maccioni, Inhibition of tau
phosphorylating protein kinase cdk5 prevents beta-amyloid-induced
neuronal death, FEBS Lett. 459 (1999) 421–426.
[22] J. Lucas, F. Hernandez, P. Gomez-Ramos, M.A. Moran, R. Hen, J. Avila,
Decreased nuclear beta-catenin, tau hyperphosphorylation and neurodegeneration in GSK-3beta conditional transgenic mice, EMBO J. 20 (2001)
27–39.
[23] G.N. Patrick, L. Zukerberg, M. Nikolic, S. de la Monte, P. Dikkes, L.H.
Tsai, Conversion of p35 to p25 deregulates Cdk5 activity and promotes
neurodegeneration, Nature 402 (1999) 615–622.
[24] M.S. Lee, Y.T. Kwon, M. Li, J. Peng, R.M. Friedlander, L.H. Tsai,
Neurotoxicity induces cleavage of p35 to p25 by calpain, Nature 405
(2000) 360–364.
[25] A. Alvarez, J.P. Munoz, R.B. Maccioni, A Cdk5–p35 stable complex is
involved in the beta-amyloid-induced deregulation of Cdk5 activity in
hippocampal neurons, Exp. Cell Res. 264 (2001) 266–274.
[26] R.A. Quintanilla, D.I. Orellana, C. Gonzalez-Billault, R.B. Maccioni,
Interleukin-6 induces Alzheimer-type phosphorylation of tau protein by
deregulating the cdk5/p35 pathway, Exp. Cell Res. 295 (2004) 245–257.
[27] Z. Mao, A. Bonni, F. Xia, M. Nadal-Vicens, M.E. Greenberg, Neuronal
activity-dependent cell survival mediated by transcription factor MEF2,
Science 286 (1999) 785–790.
[28] X. Gong, X. Tang, M. Wiedmann, X. Wang, J. Peng, D. Zheng, L.A. Blair,
J. Marshall, Z. Mao, Cdk5-mediated inhibition of the protective effects of
transcription factor MEF2 in neurotoxicity-induced apoptosis, Neuron 38
(2003) 33–46.
D.I. Orellana et al. / Biochimica et Biophysica Acta 1773 (2007) 254–263
[29] S.E. Chen, E. Gerken, Y. Zhang, M. Zhan, R.K. Mohan, A.S. Li, M.B.
Reid, Y.P. Li, Role of TNF-{alpha} signaling in regeneration of
cardiotoxin-injured muscle, Am. J. Physiol.: Cell Physiol. 289 (2005)
C1179–C1187.
[30] S. Gandy, The role of cerebral amyloid beta accumulation in common
forms of Alzheimer disease, J. Clin. Invest. 115 (2005) 1121–1129.
[31] Y.T. Kwon, A. Gupta, Y. Zhou, M. Nikolic, L.H. Tsai, Regulation of
N-cadherin-mediated adhesion by the p35-Cdk5 kinase, Curr. Biol. 10
(2000) 363–372.
[32] L.H. Tsai, I. Delalle, V.S. Caviness Jr., T. Chae, E. Harlow, p35 is a neuralspecific regulatory subunit of cyclin-dependent kinase 5, Nature 371
(1994) 419–423.
[33] S. Humbert, R. Dhavan, L. Tsai, p39 activates cdk5 in neurons, and is
associated with the actin cytoskeleton, J. Cell Sci. 113 (Pt 6) (2000)
975–983.
[34] E.M. Snyder, Y. Nong, C.G. Almeida, S. Paul, T. Moran, E.Y. Choi, A.C.
Nairn, M.W. Salter, P.J. Lombroso, G.K. Gouras, P. Greengard, Regulation
of NMDA receptor trafficking by amyloid-beta, Nat. Neurosci. 8 (2005)
1051–1058.
[35] J. Busciglio, A. Lorenzo, J. Yeh, B.A. Yankner, Beta-amyloid fibrils
induce tau phosphorylation and loss of microtubule binding, Neuron 14
(1995) 879–888.
[36] Y.L. Zheng, S. Kesavapany, M. Gravell, R.S. Hamilton, M. Schubert, N.
Amin, W. Albers, P. Grant, H.C. Pant, A Cdk5 inhibitory peptide reduces
tau hyperphosphorylation and apoptosis in neurons, EMBO J. 24 (2005)
209–220.
[37] E. Kim, F. Chen, C.C. Wang, L.E. Harrison, CDK5 is a novel regulatory
[38]
[39]
[40]
[41]
[42]
[43]
[44]
263
protein in PPARgamma ligand-induced antiproliferation, Int. J. Oncol. 28
(2006) 191–194.
T. Ishizuka, H. Ino, K. Sawa, N. Suzuki, M. Tatibana, Promoter region of
the mouse cyclin-dependent kinase 5-encoding gene, Gene 166 (1995)
267–271.
A.K. Fu, W.Y. Fu, A.K. Ng, W.W. Chien, Y.P. Ng, J.H. Wang, N.Y. Ip,
Cyclin-dependent kinase 5 phosphorylates signal transducer and activator
of transcription 3 and regulates its transcriptional activity, Proc. Natl.
Acad. Sci. U. S. A. 101 (2004) 6728–6733.
Y. Yamada, I. Kirillova, J.J. Peschon, N. Fausto, Initiation of liver growth
by tumor necrosis factor: deficient liver regeneration in mice lacking type I
tumor necrosis factor receptor, Proc. Natl. Acad. Sci. U. S. A. 94 (1997)
1441–1446.
F. Liu, I. Grundke-Iqbal, K. Iqbal, C.X. Gong, Contributions of protein
phosphatases PP1, PP2A, PP2B and PP5 to the regulation of tau
phosphorylation, Eur. J. Neurosci. 22 (2005) 1942–1950.
G. Schett, C.W. Steiner, Q. Xu, J.S. Smolen, G. Steiner, TNFalpha
mediates susceptibility to heat-induced apoptosis by protein phosphatasemediated inhibition of the HSF1/hsp70 stress response, Cell Death Differ.
10 (2003) 1126–1136.
C.C. Chao, T.W. Molitor, S. Hu, Neuroprotective role of IL-4 against
activated microglia, J. Immunol. 151 (1993) 1473–1481.
H. Takeuchi, S. Jin, J. Wang, G. Zhang, J. Kawanokuchi, R. Kuno, Y.
Sonobe, T. Mizuno, A. Suzumura, Tumor necrosis factor-alpha induces
neurotoxicity via glutamate release from hemichannels of activated
microglia in an autocrine manner, J. Biol. Chem. 281 (2006)
21362–21368.